02
02Step one
Pick a model with a free endpoint
Open the catalogue and switch on the Free Endpoint filter so you only see models you can actually call for nothing. The home page shelf is titled “Use Inference Endpoints — Free inference with leading models” and lists real ones: z-ai glm-5.2, nvidia nemotron-3-ultra-550b-a55b, deepseek-ai deepseek-v4-pro, thinkingmachines inkling.